We bring to you a 3 part interview series with the BRIDGEi2i Leadership. Second in the the series, we talk to Ashish Sharma, Director and Co-Founder at BRIDGEi2i. In this Interview, Ashish shares his views on the demand for Analytics in the financial services space and the challenges the industry is facing. He also talks about how one should go about building a career in analytics.
Here is the transcript of the interview:
INTERVIEWER: In a recent article, I read claims that Analytics is one of the most exciting career options in this century…would you agree?
ASHISH SHARMA: I would agree. I think, This is really exciting and it is turning out to be a revolution which would create the next level of competitive advantage for companies in terms of how they run their businesses and how they differentiate themselves. Given the competition, it is really exciting to be a part of what analytics can do in driving business value.
INTERVIEWER: Ashish, you have been associated with the analytics industry for more than a decade now. How do you think analytics has been absorbed in the financial domain?
ASHISH SHARMA: Financial services industry has been in the forefront of adopting analytics for several decades now, just because of the sheer volume of data available, like a lot of information on what consumers buy. I think they have been the early adopters in using data driven decision but obviously things are changing. If you think about the financial services industries today, recession was the big shift in terms of how consumers look at credit, to what extent capital is available, to what extent banks can do things based on regulatory oversight from the Fed and other government bodies. But I think it is a different field today with increasing competition, new players coming, new channels like digital media, new business models and it is an exciting world of what they can do and how differently they can do it. Obviously given their comfort in data driven decision making they are thinking what analytics can do to help them do better under new circumstances.
INTERVIEWER: If we were to compare the Indian financial services industry with that of the US, there are a lot of differences – from lack of data to lack of credit information. How different is analytics for the Indian financial services industry when compared with that of the US?
ASHISH SHARMA: That’s a great question. Having been in India and worked with financial services companies I would say it is absolutely a different ball game. You cannot bring in a solution that worked in developed markets and say, here is how you should do direct marketing because it is a very different market ball game. Is there a role data could play in having better decision in the Indian context? This is a challenging question.
In India, you don’t have consumer data for all the transactions across all the banking products or credit products that a customer may have neither we have enough demographic data of the customers compared to the US. But there is a big role that analytics could play to deal with these challenges and that is where what you do in the emerging market such as India. What you do in India is different from what you do in developed market but analytics could help take better decision even with this data concern. For example, we have engaged with customers where clients could look at customers’ information at segmented level and infer about purchase propensity that consumers could have. Depending upon which apartment complex you stay I know how well you are but not know how much you get paid, looking at what kind of life style you have, where do you go and buy what kind of vacations you take and many others. There are lots of ways of integrating the segmented information to meaningfully infer from it both from the marketing and risk perspective to take meaningful decision. I think that is the way emerging markets would have to look at it.
INTERVIEWER: We know there is lot of data around the customers that is getting generated, at one end it is good but don’t you feel that with lot of unstructured data in social media it is somewhere the institutions are getting lost on deciding which data is relevant and which data is junk? So, how do you think the institutions should decide upon correct data they need for analysis, since choosing wrong data will lead to wrong decisions.
ASHISH SHARMA: Completely agree and a great question. If you think about big data and how world has changed after McKinsey has come up with seminal report on what big data can do for their enterprises and the fact that IBM took the smarter planet as their corner stone of their marketing programs, there is a lot of awareness on what data could do but I think it has also increased the anxiety among decision makers and CIOs. We are getting into discussion where people want to know what we can do with big data. So, it has become a buzz word. Discussions are about what I should be doing in the social media, what should I be doing with location data. It is very interesting because there is hype around it.
The discussion we have with our customers is on data as means to an end by defining priorities first and the end parameter. We work backward to understand what suits our client best. And which is why Analytics is never a one stop game or a one day wonder but companies that want to differentiate and outperform competition will make data driven decision a way to run their businesses. In that, processes deciding on what data streams are valuable and how you bring them together are all enablers.
But it starts with the fundamental question what I want to achieve? Let’s say I think from marketing perspective – Suppose I have a great product and great customer base and I want to expand and I have another 15 products that I want to sell but don’t know who to sell to and I want to drive my product penetration. The question is, how can I know what the customer needs? And if you start with that objective then you can start thinking about how social media can help understand customer need through customer experience data, website navigation pattern can tell you something about what the customer is looking for.
So, I think you should define the need first, define the end objective and then come back. It is not a one day journey; you won’t get the right answer in the first instance. You have to start with some priority in mind and work from there what value you can generate.
INTERVIEWER: I recently came across a survey where a majority of the managers accept that analytics has its place in decision making but it is also true that there has been inertia in accepting analytics as a part of day to day activity. What do you think would be the first step for an organization looking to implement analytics towards decision making?
ASHISH SHARMA: The corner stone of what we do at BRIDGEi2i is impact. We wanted to make sure that it is not lost in what we are trying to do and that is how we defined our company saying BRIDGEi2i bridging information to impact. As i said, if you are doing analytics as a one-time activity or as a project it won’t lead to sustainable competitive advantage and that is where the difference comes. Companies able to adopt analytics and make it corner stone of decision making. Think about capital one, they made a whole new business about what can be done and one of the things we are focusing on is how do you make analytics consumable.
Ultimately, Analytics is not for the sake of doing analytics. It is you as a decision maker being combative, should say this is useful and I’ll take decisions based on this. And we are doing couple of things around it. I think companies should focus on making insight visible and actionable in real time. And there is where we are building couple of solutions, they are on the fly solution, they could be reside on the cloud- around visualization. Interesting ways to look around portfolio, interesting ways to manage model, interesting ways to understand customers behavior etc. Analytics gives decision makers comfort on what’s happening and it has to be interesting because it is not traditional BI not backward looking it has to be forward looking and actionable. It is one big area of focus. All the intelligence, predictive stuff has to be baked in and ultimately translate to something manager could use.
We are not off the shelf product company but we are a focused solution company. It is not like can we do this for you but rather can you build a solution for you that’s allows your managers to do more and consume more of analytics and data driven insight on an ongoing basis. We built solution for them which allow them to adopt it. As companies think today, I’ll say two things, one start with what is the value and how can you make the human element consume analytics is becoming more important than the analytics. And that’s what should take precedence and dictate all of what to do from there on. I will do whatever it takes for my portfolio marketing manager to be comfortable in taking data driven decision. Let decide what that is, what he needs and how he want to be more effective.
INTERVIEWER: There has been a buzz that what was IT sector in 1990s, it is analytics in 2010s. Is India a low cost game or talent driven game?
ASHISH SHARMA: I think cost is an element but just an element. Analytics is so fundamental to impact that when you start thinking it is about ROI not about cost, it’s about the quality, it’s about whether things get implemented, what will get implemented, are the solutions sustainable or there are dependencies; say I have the best analyst in my team, I have got the best decision support mechanism but the moment I don’t have him/her, then what do I do. So, as an analytics company you got to be effective, drive business impact, make sure things work, get implemented and there is a feedback loop to learn what works and what doesn’t. Compared to the value analytics can create, cost is miniscule. But if you don’t have the value then it is irrelevant to have cost advantage.
INTERVIEWER: The banking industry is faced with many challenges; on one side RBI takes stringent measures to suck the liquidity from the system. On the other hand NPAs have also doubled over the past few years. All this affects a bank’s profitability. At the same time we have also seen increased analytics adoption by banks. I think there is a contradiction here…why does this happen?
ASHISH SHARMA: I think the context is different between US, UK, the developed market, and the emerging market. In Post-recession US, the regulatory authorities have significantly changed the goals of the game. It is a very different playing field. With change in economic circumstances the credit appetite of consumer has gone down and at some point we saw delinquencies rise.
From analytics and decision making perspective you need to step into chief marketing officer’s shoes. They are thinking about, how I would play this new game, how can I reach out to customers when fair credit act will have 15 different constraint on what can be done and what can’t. How do I take the right lending decision while making sure the regulators are comfortable on how I am granting credit, how am I running my business from risk and capital perspective. So, there is a role analytics could play.
It is a very different play than what was a few years ago. There is lot of focus on regulatory compliances, lot of focus on pricing decision based on compliance with the new regulatory practices. This has changed the room companies have and hence they need to be smarter to be able to drive profitability.
In Indian context it is different. Economic growth has slowed down in the recent quarter. You suddenly see the delinquencies increase and on the other hand you see growth become tougher. It is a double whamming for them. We have seen a lot of interaction with the Indian banks now. Most strategic interaction saying in this circumstances are how can I drive growth? How do I classify my customer better?
INTERVIEWER: What are the challenges, related to finance domain, in model building? Say, building statistical models for housing loan and insurance policy etc.
ASHISH SHARMA: Challenges vary depending on context. Model building is easy but the question to ask is what are you modelling for? Many times when things get wrong, it is not because of the model but you started in the wrong direction, you started to model the wrong phenomenon. I will say that’s the fundamental and that where we spend a lot of time with our customer, thinking through how this will be implemented, what kind of decision it will support, what kind of information you can leverage on when we do this.
Models can go wrong because you haven’t thought through question like what data stream could be valuable as you take this decision. Models can go wrong because you haven’t considered of end implementation in your mind. Models can be weak because you haven’t brought all the information to the table. Those are some of the generic level challenges around model development.
Financial Service Company have been using model for several decades to differentiate. Generic models no longer work because the world is moving towards a more personalized view. Say, you want something personalized to yourself than to personalize to broader categories of customers. Hence, techniques which are applicable in those context being more micro, more personalized, more customized. We are talking about very different stream of analytical model development than what used to be the realm earlier. Make sure that you pick up the right technique, make it more granular, pick the right frame work, data stream are some of those key element I would think of.
INTERVIEWER: Does this (customized model in finance domain) also mean that life cycle of the model itself is very short compared to other industries? What do you feel?
ASHISH SHARMA: Yes. We are seeing a confluence of statistician who use to come from probabilistic world and think about samples of phenomenon, how do you infer what happens, to computer scientist who can think about algorithm which can self-learn, self-detect pattern. You always want to be right. So, you always want to say my model is performing in the way it is supposed to perform. Earlier, you could start a debate on how long the model can be accurate. Now the discussion is changing to: Why can’t I keep tweaking my model on a more frequent basis. It’s a good debate on the right question. I think it’s a question between accuracy and interpretation; regulators also want comfort.
So, you just can’t have something which is a black box which the regulator can’t understand versus being right and accurate. Yes people are refining their model more frequently; they are trying to think about technique that can do that in some form fairly more relatively more frequently but it’s also a question of getting comfort around why my model is getting refined. So, while the time might have become shorter, you refine a model more frequently but you also have to think about the flip side i.e. Why are they getting refined more frequently? What is changing the customer preferences that drive this change?
INTERVIEWER: There must be efficient way to track all this models and we as an analytics organization must do something about it. What’s your thought on those fronts?
ASHISH SHARMA: Absolutely, it is becoming a regulatory focus. We are working with companies to define solution in that realm, we are building frame work where companies could look at model risk and manage model risk. That’s an area where we have strong capabilities and we are working with customers to see how we could be of help in that context to them. So, its big area in terms of regulators is focusing on it from model risk prospective. We got a very good frame work that we currently partnering with clients to see how they could benefit and address some of those concerns.
INTERVIEWER: How has your journey been so far with BRIDGEi2i?
ASHISH SHARMA: Fantastic. It is a great team and that’s the fun part. It is a great area to be in with the expertise in the area where we chose to be in which is financial services and insurance, technology and trying to help these companies become better in taking decision. So, I think we have had fantastic progress. We have got our two pillars from technology prospective, Surveyi2i in the market, and we are commercializing it. We have couple of other solutions where we are partnering with clients. We are working on fundamentally different newer unstructured problem solving for clients. We are engaging with client for longer term for data driven decision and value creations. We are very positive on where we can partner with clients and it just keeps us excited and focused on what we should be doing and where we can create value for enterprise over the period of time.
INTERVIEWER: What would be your suggestion to those who want to enter into the space of analytics? What steps they need to take to enter into the world of analytics?
ASHISH SHARMA: Certainly, the right question. There are different ways the question get asked, the other way the question get asked is, should I be a data scientist, how can be I a data scientist. There are buzz word created around like, who are this guys is that blue collar worker this are data scientist; it is the new working categories that we sought of created. It is great space to be in, but I think it needs a very different set of skills, dedication and commitment to excel in what you could do in this space. I agree with the wow and wow is what excites us to be here but there is a lot of rolling up of your sleeves to build the skills for your future as this is the space that is changing as we speak. What we do today was not something that we did 15 years before.
You think about this, we are doing new things today and I know as we speak there are lots of companies investing in newer technology, new business model coming in and new ways to consume data. Think about Surveyi2i, who could have thought one can get on the fly ability to analyze and get insight into unstructured data on any survey data you put it. So things are changing it is not just us.
To your question, if you want to be in this space I will say a couple of things, one has to be very comfortable with being quantitative in nature. It is not just about statistics, machine learning and computer science but being able to quantify matters a lot. You will miss the big picture if you just detail it down to means or techniques. I have seen great analysts and we have several of them in our team.
I think how can I hire more such guys and the answer is I not have to hire statistician but I have to hire and get people on board who are inherently quantitatively inquisitive. I think that’s the corner stone of a good data scientist. This is not an area or stream where you can give just advice but here you consult based on the facts. So you have to very comfortable getting in to the details. As an analyst, I love doing that and that’s the only way to build your career in this industry. Two things important, first: one has to be inherently quantitatively inquisitive and secondly: very comfortable being detail oriented, rolling up your sleeves and diving deeper into what striving this how can I validate this. When you can ask the right kind of question to yourself then you have the capability to answer.
BRIDGEi2i provides Business Analytics Solutions to enterprises globally, enabling them to achieve accelerated business impact harnessing the power of data. Our analytics services and technology solutions enable business managers to consume more meaningful information from big data, generate actionable insights from complex business problems and make data driven decisions across pan-enterprise processes to create sustainable business impact. To know more visit www.bridgei2i.com
The views and opinions expressed in this article are those of the author and do not necessarily reflect the official position or viewpoint of BRIDGEi2i.